Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Transformers Can Learn Posterior Predictive Distributions In-Context
Gyeonghun Kang, Changwoo J. Lee, Xiang Cheng
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add attention normalization to the transformer (ablation) | GP regression PFN transformer (L=32, C=256) · RBF regression n-generalization (d=8) | TV distance | Improved |
| increase attention depth L (ablation) | GP regression PFN transformer · BLR and RBF regression | TV distance | Improved |
| increase output bin resolution C (ablation) | GP regression PFN transformer · BLR and RBF regression | TV distance | Improved |
| increase pretraining range n_max and depth L (ablation) | normalized PFN transformer · RBF regression n-generalization (d=16) | prediction error, coverage, TV | Improved |
| increase evaluation sample size n' beyond pretraining range at fixed L and n_max (ablation) | finite-iteration solver transformer · RBF regression (d=16) | solver error | Hurt |
Closest papers by meaning
- Can Transformers Learn Full Bayesian Inference in Context? (ICML 2025 · similarity 0.89)
- Position: The Future of Bayesian Prediction Is Prior-Fitted (ICML 2025 · similarity 0.88)
- Bayesian Neural Scaling Law Extrapolation with Prior-Data Fitted Networks (ICML 2025 · similarity 0.87)
- In-context Convergence of Transformers (ICML 2024 · similarity 0.86)
- Streamlining Prediction in Bayesian Deep Learning (ICLR 2025 · similarity 0.86)
- Nonlinear transformers can perform inference-time feature learning (ICML 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
